Chapter 7
Stack and Subroutines
Updated: 3/23/15
Basic Idea
l Large programs are hard to handle
l We can break them to smaller programs
l They are called subroutines
l Subroutines are called from the main
program
l Writing subroutines
l When should we jump? (use CALL)
l Where do we return to? (use RETURN)
Subroutine
l A subroutine is a block of code that is called from different
places from within a main program or other subroutines.
l Saves code space in that the subroutine code does not have to
be repeated in the program areas that need it;
l Only the code for the subroutine call is repeated.
l A subroutine can have
l parameters that control its operation
l local variables for computation.
l A subroutine may pass a return value back to the caller.
l Space in data memory must be reserved for parameters, local
variables, and the return value.
Subroutine
Using Subroutines
l Whenusing subroutines we need to
know the following:
l Where is the NEXT instruction’s address
l How to remember the RETURN address
l Subroutines are based on MPU
instructions and use STACK
Stack
l Temporary memory
storage space used STKPTR
during the execution of a
program
l Used by MPU
l Stack Pointer (SP) PC
l The MPU uses a register
called the stack pointer,
similar to the program
counter (PC), to keep track
of available stack locations.
Data Storage via the Stack
l The word ‘stack’ is used because storage/retrieval of
words in the stack memory area is the same as
accessing items from a stack of items.
l Visualize a stack of boxes. To build a stack, you place
box A, then box B, then box C
l Notice that you only have access to the last item placed on the stack (the Top
of Stack –TOS). You retrieve the boxes from the stack in reverse order (C
then B then A). A stack is also called a LIFO (last-in-first-out) buffer (similar to
a Queue)
PIC18 Microcontroller Stack
Content:
l Consists of 31 registers-21- TOSU/H/L
bit wide, called the hardware
stack 21-bit
Address:
STKPTR
l Starting with 1 to 31 (5-bit)
l Stack is neither a part of
program memory or data
registers.
l To identify these 31 registers,
5-bit address is needed
l PIC18 uses one of the special
function registers called
STKPTR (Stack Pointer) to
keep track of the available
stack locations (registers).
STKPTR (Stack Pointer) Register
l SP4-SP0: Stack Address
l STKOF: Stack overflow
l When the user attempts to use more than 31 registers to
store information (data bytes) on the stack, BIT7 in the
STKPTR register is set to indicate an overflow.
l STKUNF: Stack underflow
l When the user attempts to retrieve more information than
what is stored previously on the stack, BIT6 in the
STKPTR register is set to indicate an underflow.
Instructions to Store and Retrieve
Information from the Stack
l PUSH
l Increment the memory address in the stack pointer (by
one) and stores the contents of the counter (PC+2) on the
top of the stack
l POP
l Discards the address of the top of the stack and
decrement the stack pointer by one
l The contents of the stack (21-bit address), pointed
by the stack pointer, are copied into three special
function registers
l TOSU (Top-of-Stack Upper), TOSH (High), and TOSL (Low)
TOSU TOSH TOSL
Instructions to Store and Retrieve Information
from the Stack
l The PIC18 stack has limited capability
compared to other µPs. It resides within its
memory, and is limited to 31 locations.
l For a CALL, address of next instruction
(nPC) is pushed onto the stack 0: left alone!
l A push means to increment STKPTR, then
store nPC (Next PC or PC+2) at location
[STKPTR].
l STKPTR++; [STKPTR] ←nPC
l A return instruction pops the PC off the
stack.
l A pop means read [STKPTR] and store to the
PC, then decrement
l STKPTR (PC ←[STKPTR], STKPTR--)
Example
l What
is the value of PC, TOSU/H/L and
STKPTR as you execute each line?
nPC TOS STKPTR W
22 0 0 00
24 0 0 20
26 26 1 20
28 28 2 20
2A 26 1 20
2C 0 0 20
Subroutine Call
l In the PIC18F, the stack is used to store the
return address of a subroutine call.
l The return address is the place in the calling
program that is returned to when subroutine
exits.
l On the PIC18Fxx, the return address is PC+4,
if PC is the location of the call instruction.
l The return address is PC+2 if it is a rcall
instruction.
Call and Return Instructions (1 of 3)
l CALL Label, S (0/1) ;Call subroutine
Remember: CALL is
a 2-word Instruction!
; located at Label
l CALL Label, FAST ;FAST is equivalent to
;S=1
l If S = 0: Increment the stack pointer and store the contents
of the program counter (PC+4) on the top of the stack (TOS)
and branch to the subroutine address located at Label.
l If S = 1: Increment the stack pointer and store the contents of
the program counter (PC+4) on the top of the stack (TOS) and
the contents of W, STATUS, and BSR registers in their
respective shadow registers, and branch to the subroutine
address located at Label.
RETURN
l RETURN,0 à gets the address from TOS and
moves it to PC, decrements stack pointer
l RETURN,1 à gets the address from TOS and
moves it to PC, decrements stack pointer;
retrieves all shadow registers (WREG,
STATUS, BSR)*
l RETLW à gets the address from TOS and
moves it to PC ; returns literal to WREG,
decrements stack pointer
* 1 or FAST
Call and Return Instructions (2 of 3)
l RCALL, n ;Relative call to subroutine
within n = ± 512 ;words (or ± 1 Kbyte)
;Increments the stack pointer and stores the
contents of the program counter (PC+2) on
the top of the stack (TOS) and branch to the
location Label within n = ± 512 words (or ±
1 ;Kbyte)
Example
l Program Listing with Memory Addresses
Org 0x40
2-Word Instructions
END ;will be at the end of the program!
After CALL: After RETURN:
Note:
2-Word TOS=00 00 2E
àInst. PC=00 00 40 PC=2E
PC+4 (2Aà2E)
STKPTR=01 STKPTR=00
Subroutine Architecture
How do we write a subroutine?
Parameter Passing
Inputs Basic Functionality
Outputs
Register Modifications
List of Subroutines used
Macros and Software Stack
l Macro
l Group of assembly language instructions that can be
labeled with name
l Short cut provided by assembler
l Format includes three parts
Push_macro macro arg
movff arg,POSTINC1
endm
USE: Push_macro WREG
A push means to increment STKPTR,
then store nPC (Next PC or PC+2) at
location [STKPTR].
STKPTR++; [STKPTR] ←nPC
Macro Description - Example
l See how FSR is loaded and POSTDEC works.
l How a MACRO is being called!
Before MAIN
MACRO Application
l Note COUNT is not defined in the MARCO
l It is the "arg” of the MACRO
l MACRO is assembled after every instance it is called
MACRO Application
l So what if MACRO is called multiple times?
l A MACRO is assembled after every instance it is
called
Subroutine versus Macro
l Subroutine (by MPU) l Macro (by assembler)
l Requires instructions l Based on assembler
such as CALL and
l Shortcut in writing
RETURN, and the
STACK (overhead) assembly code
l Memory space required l Memory space
by a subroutine does required depends on
not depend on how how many times it is
many times it is called called
l It is less efficient in
l In terms of execution
terms of execution than it is more efficient
that of a macro because
it includes overhead because it does not
instructions such as have overhead
Call and Return instructions
More about subroutines…
l Remember subroutines can call other
subroutines
l This is referred as structured code
Using Table Pointers
l Reading/writing
values from/into the
program memory one byte at a time
TBLWT* / TBLWT*+ / TBLWT*-
Table Latch (8-bit)
Write into Memory
TBLPTRU/H/L (21-bit) Word
Read from the Memory
Program
Memory Table Latch (8-bit)
(16-bit) TBLRD* / TBLRD*+ / TBLRD*-
Table Example
Rd the register content
into Table Latch
Save in Prog Memory; Label the value as BUFFER
TBLPTR(U/H/L)= 00 00 40
0x000040 0x0002 TABLAT=0x02
Pointing to the address
Program
Memory W=0x02
(16-bit)
REG10=0x02
Table Example LAB: Modify this program such that
values 0xaa, 0xbb, 00cc stored in the
program memory are copied into
REG60,61,62, respectively
TBLPTR = 00 00 40
0x000040 0x0002 TABLAT=0x02
Pointing to the address
Program
Memory W=0x02
(16-bit)
REG10=0x02
Examine this code:
See next slide…..
Answer the following:
l At what location in the program memory CLEARME is built? Explain.
l What are the contents of register 0x60, 0x61, etc. in the program memory?
l Where is the location of FSR1 when the MARCO is called?
l Where exactly does CLEARME macro does? How many registers are effected?
l Where exactly does BYTECP macro does? How many registers are effected?
l Modify the program using MACROs such that you perform the following tasks:
l Copy NINE values already stored in PROGRAM MEMORY locations, starting with locations
0x80, into RAM locations starting with register 0x80. Assume the numbers are 1-9.
l Copy NINE values already stored in PROGRAM MEMORY locations, starting with locations
0x80, into RAM locations starting with register 0x90. Assume the numbers are 1-9.
l Take the sum of all the values and locate the SUM in RAM location 0x100.
l Delete all the RAM registers starting with with register 0x90 - 0x09F
Example…..
l Modify program 7.5.3 such that you can
correctly take the average of any sum.
l The number of inputs can be up to 20
non-zero unsigned values
Unsigned Division Operation
l The PlCI8 MCU does not provide any divide
instruction. Therefore, a divide operation must be
synthesized by other instructions. A simple but popular
divide algorithm in use today is the repeated
subtraction method. This method performs unsigned
divide operation. The hardware required for
implementing the repeated subtraction method is
shown in Figure below
l Before performing the repeated subtraction operation,
one needs to load 0, the dividend, and the divisor info
registers R, 0, and N, respectively. The carry flag is
used to indicate whether the subtraction result is
negative. The ALU can perform n-bit unsigned addition
and subtraction operations. The repeated subtraction
method consists of n steps. Each division step consists
of three parts:
l Step 1
l Shift the register pair (R, 0) one place to the left.
l Step 2
l Subtract the contents of N from R and put the result back
in R if the result is positive.
l Step 3
l If the result of Step 2 is negative, then set the least
significant bit of 0 to o. Otherwise, set the least significant
bit of 0 to 1.
Reference: Huang
Division By Subtraction
References
l http://www.ece.msstate.edu/~reese/
ece3724/lectures/chap6_subr_ptrs.pdf
Subroutine Documentation and
Parameter Passing
l Parameter passing
l Information exchanged between a calling program and a
subroutine
l Subroutine documentation should include:
l Function of a subroutine: Brief description of what it does
l Input parameters: Information that should be provided by
calling program to subroutine
l Output parameters: Information or results provided by
subroutine to calling program
l Registers modified: List of registers changed by the
subroutine
l List of subroutines called: List of other subroutines called by
the called subroutine
